On Tue, 16 Jun 2020 at 11:40, Philippe Mathieu-Daudé <f4...@amsat.org> wrote: > On 6/16/20 12:27 PM, Peter Maydell wrote: > > This change seems kind of pointless unless these GPIO lines are > > actually wired up to something. > > Yes, I should have kept it out of this series, or documented > better the goal in the cover. > > I'm setting the roots to motivate a team of developers to > work on a visualization of the MPS2 board. The push-button is > supported by Zephyr, so the the idea is the visualizer generates > QMP GPIO event to be processed such in pca9552_set_led(), and > interact with the guest firmware.
I think that having a framework so we can better model this kind of push button / LED / similar thing is definitely good. I just think we need to review it at the framework level first -- it might turn out that actually the right way to wire up the push button to the UI framework isn't with a GPIO wire at all. Similarly with the other patchset that sends QMP events for LEDs -- that also seems like it's half of a design and a bit awkward to review without the context for what it connects to. thanks -- PMM
